Conference item icon

Conference item

Partially observable stochastic games with neural perception mechanisms

Abstract:
Stochastic games are a well established model for multi-agent sequential decision making under uncertainty. In practical applications, though, agents often have only partial observability of their environment. Furthermore, agents increasingly perceive their environment using data-driven approaches such as neural networks trained on continuous data. We propose the model of neuro-symbolic partially-observable stochastic games (NS-POSGs), a variant of continuous-space concurrent stochastic games that explicitly incorporates neural perception mechanisms. We focus on a one-sided setting with a partially-informed agent using discrete, data-driven observations and another, fully-informed agent. We present a new method, called one-sided NS-HSVI, for approximate solution of one-sided NS-POSGs, which exploits the piecewise constant structure of the model. Using neural network pre-image analysis to construct finite polyhedral representations and particle-based representations for beliefs, we implement our approach and illustrate its practical applicability to the analysis of pedestrian-vehicle and pursuit-evasion scenarios.
Publication status:
Published
Peer review status:
Peer reviewed

Actions


Access Document


Publisher copy:
10.1007/978-3-031-71162-6_19

Authors


More by this author
Institution:
University of Oxford
Division:
MPLS
Department:
Computer Science
Role:
Author
More by this author
Institution:
University of Oxford
Division:
MPLS
Department:
Computer Science
Role:
Author
More by this author
Institution:
University of Oxford
Division:
MPLS
Department:
Computer Science
Role:
Author
More by this author
Institution:
University of Oxford
Division:
MPLS
Department:
Computer Science
Role:
Author
More by this author
Institution:
University of Oxford
Division:
MPLS
Department:
Computer Science
Oxford college:
Trinity College
Role:
Author
ORCID:
0000-0001-9022-7599


Publisher:
Springer
Host title:
Formal Methods. FM 2024
Pages:
363–380
Series:
Lecture Notes in Computer Science
Series number:
14933
Publication date:
2024-09-11
Acceptance date:
2024-07-10
Event title:
26th International Symposium on Formal Methods (FM'24)
Event location:
Milan
Event website:
https://www.fm24.polimi.it/
Event start date:
2024-09-09
Event end date:
2024-09-13
DOI:
EISSN:
1611-3349
ISSN:
0302-9743
EISBN:
978-3-031-71162-6
ISBN:
978-3-031-71161-9


Language:
English
Pubs id:
2011377
Local pid:
pubs:2011377
Deposit date:
2024-07-01

Terms of use



Views and Downloads






If you are the owner of this record, you can report an update to it here: Report update to this record

TO TOP